An application of the remote sensing derived indices for drought monitoring in a dry zone district, in tropical island

Abstract

Recent research has shown that droughts have intensified in South Asia over the past two decades. As a natural disaster, this has severely impacted people’s livelihoods, especially in the dry zones of Sri Lanka. Thus, to ensure human well-being, security of water resources, and ecosystem health, it is essential to minimize the impacts of droughts using reliable information. Remote sensing (RS) data and techniques help bridge the gap by enabling the analysis of drought phenomena through a diverse array of indices developed in the fields. However, until now, there has been a lack of systematic monitoring and reliable data for accurately characterizing droughts in the study area. As the first comprehensive analysis, we tried to evaluate the spatial–temporal dynamics of drought conditions in two Divisional Secretariat Divisions of the Anuradhapura district, Sri Lanka using eight standardized remote sensing-derived indices over a decade (2013–2023) including the Standardized Precipitation Index (SPI). SPI values indicated that the region has experienced notably below-average precipitation. According to SPI results, a significant portion of the Medawchchiya area experienced arid conditions in 2023. All other indices proved that 2018 was the driest year and 2013 was the wettest year among the three time points, as reflected by their low and high index values. However, according to NVSMI and LST, the wettest year is 2023, with only 1.78 % of areas experiencing severe drought and a maximum LST of 31.4 °C. LULC change detection revealed that 14.3 % of agricultural lands and 5.1 % of forest areas were converted into barren lands over a decade. Overall, this conversion may be another leading factor contributing to increased LST and dryness in the area during the concerned period while increasing the mean LST of barren land by around 4.7 °C per decade. Land surface-related and vegetation-related indices, such as NDWI, NDMI, LST, and NDVI, exhibited a more pronounced impact on short-term drought occurrences. The findings revealed that average precipitation coincides with short-term drought episodes in the area, with 2018 standing out as having the least rainfall and the driest year. The study’s findings may provide additional insights for planning authorities, supporting environmental protection and enhancing agricultural production by mitigating droughts’ impacts through short- and long-term strategies. Although, the study focused on a small area, a similar approach could be extended to other areas by incorporating advanced machine learning techniques and additional drought indices in the future.